Linux as Wintel parasite
Summary: When Linus Torvalds choose to tie his new kernel to x86 he did so thinking, quite correctly as it turned out, that Microsoft's tendency to software bloat would provide an unending stream of cheep hardware for Linux developers and users. Great! except there's a side effect: Linux is ultimately dependent on Microsoft because its raison d'etre f is almost literally to live on Microsoft's leavings.
Last week Linus Torvalds told a seminar group at Portland's LinuxCon that Linux is getting a little bloated - a consequence of the big blob kernel architecture required by his decision to prefer the efficiency of directly using x86 interrupts to the much more hardware independent architecture Tannenbaum developed Minix to teach.
Sun blogger Joerg Moellenkamp said something particularly interesting about this:
Of course it's a nice sign of success, when people port more and more stuff to an operating environment and into the kernel. Perhaps this is the price of success. But at foremost it's a problem. Bloat isn't just about using more memory, it's about speed as well.The Register delivers another interesting piece of information:
Citing an internal Intel study that tracked kernel releases, Bottomley said Linux performance had dropped about two per centage points at every release, for a cumulative drop of about 12 per cent over the last ten releases....
Should they rearchitect Linux for the future (the SunOS/Solaris moment for the Linux community). And as refactoring, optimization and rearchitecting are tedious and boring tasks: Who will do it? I think, the next few years will be interesting ones for Linux.
Another speaker at the same event, IBM's Bob Sutter, really needs to spend a few minutes looking at the history of his own company's VM product line, but other than that came up with another absolute shocker: Linux won't succeed on the desktop, he said, unless it creates a unique Linux desktop - or, in translation, that Linux can't lead by following.
Personally I think that the SuSe business desktop does lead Microsoft in some areas, but, of course, Mr. Sutter wants to sell cloud computing - and so does Eric Mandel, CEO of a company called Blackmesh, providing managed Linux hosting services. He does a very sad and funny presentation on doing what they did: implementing a couple of open source deployment tools (Puppet and Cobbler) to make it fairly easy to configure and deploy Linux server/application combinations. This can be very important in their business, but I thought the retrograde nature of both the solution and its markets unhappily captured the essence of Linux today.
Basically he's using an open source evolution of the old Jumpstart stuff to provision gear for customers who haven't figured out yet that letting other people control both their data and their most critical business infrastructure is a recipe for coming to a quick and unhappy end. Cool stuff, for five years ago - but completely obsoleted for customers by today's cost/risk trade-offs in doing it themselves and for techies by Solaris zones.
When you look at this kind of thing the contrast with BSD could hardly be greater. That group's focus, despite their many divergences and disagreements, is always on better, faster, smaller - and Apple's posture as the anti-IBM in personal computing carries over to its relationship with the BSD community: it's the world's biggest producer of Unix personal computers, but it doesn't try to direct BSD research and it hasn't tried to build services revenues on its own limitations.
Mr. Torvalds set out to build a "free Unix for the 386" and succeeded brilliantly in doing so - but both its internal architecture and its market success depend on the peculiar dynamics of the wintel market in which x86 forms the common ground between the huge majority using Microsoft software and a rebel group looking for something to call its own.
Thus looking at it as an outsider, I'd say that much of what made headlines at Linuxcon 2009 was in one way or the other about the chickens associated with the reinvention of old technologies for commercial gain starting homeward -with all of it demonstrating that if the Linux community didn't have Microsoft both to be against and to prop up their shared x86 foundations, it'd wouldn't exist.
And that's sad - but not irretrievable because at this point it's fundamentally a leadership failure, not a community failure, and therefore something that could be changed.
Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.
Talkback
You have some questionable ideas.
I think to even use the word 'parasite' is unfortunate, and you deserve some flack.
Secondly, had you considered how all the standard C-library functions have now been deprecated at least so far as security is concerned on public-facing sites? That might be why there has been a 2% slowdown, as more involved paranoid routines must replace the standard functions.
Thirdly, what has the bit about Blackmesh and retrograde deployments got to do with all this? If you are enlightened you can do good deployments.
Fourthly, Apple's use of BSD is a special case because they DON'T have to make it install on 'any hardware' - just their own. They have a massive advantage in terms of size and perhaps also in tuning.
Fifthly, it is insulting to companies like Novell or Red Hat to use the expression "building services revenues on its own limitations". Or were you meaning IBM? Would Sun support charges be any different?
oh?
Torvald's decision to go fit his kernel to x86 came from his perception that x86 hardware was cheap and plentiful - a consequence of the MS/IBM monopoly wars and software strategies.
Had no monopoly developed, there might well have been many more CPUs in use and that would have made the minix (OO micro-kernel) architecture the right "pragmatic" choice.
2 - the c-libs? Nothing's more secure (or faster) than OpenBSD and/or Solaris.
3 - "Thirdly, what has the bit about Blackmesh and retrograde deployments got to do with all this? If you are enlightened you can do good deployments."
Sure - but the customers buying this aren't thinking long term and the technology he uses is an extension of something the Solaris community has left behind. The point here is that it's good stuff, but for 2003/4, not 2009/10.
4 - is Apple special? Sure - but the point is that it's bigger than Linux and non intrusive.
5 - "Fifthly, it is insulting to companies like Novell or Red Hat to use the expression "building services revenues on its own limitations". Or were you meaning IBM?"
Why is it insulting? isn't this what all three do?
6 - "Would Sun support charges be any different?"
Yes - Solaris support charges are far less than those for Linux and are actually for support - not a wink nudge licensing scheme.
Rudy step away from the crack pipe!
Did you see how Red Hat's recent quarterly results ROCKED!?, making good profits in the recession and we all saw how Sun's results SUCKED!:-P
Per Tannenbaum:
I'm not sure this applies. While no one likes crashes they have become quite rare. I'm not sure when one of my machines crashed last. Especially from a driver. Power supplies fail and hard drives fail but driver architecture is such that failures are increasingly rare. And the part about features and performance could not be further from the truth.
I don't know whether microkernels and interrupt driven architecture are mutually exclusive but the interrupt driven architecture is much more efficient and provides better performance than the alternatives such as polling.
Perhaps some older hardware could be removed from the kernel. Or better yet, the microkernel be used with interrupts in an effort to decrease bloat.
Is the interrupt driven architect and microkernel mutually exclusive?
No. Yes. ok: sort of
The issue in re Linux bloat today is a consequence of over adaptation to x86 circa 1989 (or, actually, 1978). For the 80386 tying a furball to the hardware made technical (performance) sense - and, commerically, MS business practices guaranteed a supply of "obsolete" x86 hardware so the decision to do this made both technical and business sense. Now, however, they're reaching (over-reaching?) the limits of this strategy - and that's the problem.
RE: Linux as Wintel parasite
Linux isn't a company and a corporations interest depends on what implementation is profitable. No leadership? Of course there is otherwise Linux would succeed in so many areas as it does. It might not however be as prominent in the business desktop as you wish. Once again: why this obsession about a desktop when it's very doubtable that our current perception of a desktop is efficient and smart? That's actually I understand Bob Sutter. And after all quality hasn't formed the market and users are to used to awkward computing that they wouldn't anyway appreciate something better that easily. Some does but far from all. There have been many inventions over time, in different fields, superior to generally used technology, but without success. Something else is forming markets, isn't that true?
Agreed
RE: Linux as Wintel parasite
well, i agree that MS HAS the desktop market rigth into his pockets, why? 'cause ms has "learn" from his past errors (finally improving) and as murph says, users don't care a damn 'bout which OS are running behind their screens (apps), and this makes the next big challenge for all the desktop players: interoperatibilty.
if cloud computing has teach us something, that would be interoperabilty, truly client-side-OS-independency, as jason perlow has already pointed out (http://blogs.zdnet.com/perlow/?p=11167)
so, in this tendence, users, will adapt and work in browsers, not desktops, and even if that's so many ligth years away from now, take MS office to the mac, and you will have a "happy" user, take office to the linux and you still have a happy user, (or wichever app you would like to point about).
BWT I'm - by no way-, MS fanboy.
so, there is really some point is discussing 2% performance drops, when the average users, had a double, triple or quad core desktops? when the "cloudadmins" will be keeping their processing power up to date?. the average user just wants to sit, use and experience a 0 software downtime (even if ther are doing stupid or nonsese use of their pc, sorry i mean, desktop), it is we, the geek-oriented sysadmins,netadmins or wichever admin u like, are the ones who care for this kind of issues, not the average users.
want some more? let's take the finalcial side of the equation, if i can stop paying 5000 windows licences every time a new version of the OS has rolled out, and remain faithfully paying overpriced autodesk licences (let's say autocad 20xx), in a mac or in a linux desktop, damn! that will be make my entire director board really happy, and even get some extra money to more infraestructure,
So, sorry murphy for the off topic chatter, the same idea applies to the x86 interrupts or the minix type kernel, truth is if as long as ANY kernel let me keep running my apache webserver (for example), i don't really care wich type of kernel is liying below, even if that kernel is win32. (and by god i dislike MS so damn much).
ps. srry 4 the bad english..
A matter of perspective
When we're discussing what users want, we have to understand that they don't care how we deliver, only that we deliver.
In this instance, however, I'm talking about how we deliver and in a technical, not a user service, context - and in that context 12% performance loses (and the register's problems with aritmetic) matter.
RE: Linux as Wintel parasite
Any supporters left?
Now he's picking on the less than 1% of users who dare to use an open source OS. The less than 1% says it all really, no need to rub it in. As a solid, but simple OS it has come into wide use for embedded devices, simple web servers and even supercomputers. However, when it comes to acting as a global desktop, it lacks the research, development and quality systems of the proprietary model provided by MS and to a much lesser extent Apple. Open source software can make bad to adequate copies of existing proprietary software and occasionally act as a seed for proprietary companies (Sun OS anyone?), but it's generally going to be trailing edge. This isn't really a problem if you realise that Linux may be perfect for specific applications or as a safety net, but is not a global contender.
So what's left? The holy BSD apparently. I'm occasionally nostalgic for the 1960s and 70s too, but you're proposing Unix as the perfect OS?
Well at least finally you're back on topic ;-)
Where do you get your information? Microsoft?
For web servers on the Internet, according to Netcraft as of September, Apache has over 46% marketshare, compared to Microsoft's less than 22%.
In supercomputers, Linux has not "come into wide use for...supercomputers", it's the defacto gold standard: 88.6% for Linux as of June 2009 according to Top500.org, versus 1.00% for your god, Microsoft.
I'm surprised Rudy didn't call your contention that Sun OS was "seeded" by open source software. Those of us who know history (without Microsoft's influence) know that Sun's original Unix software was based on the portable Version 7 Unix, not BSD. They got their version from Unisoft back in '82. Freely distributable BSD as we now know it didn't come about until 1991, with the release of 4.3BSD and Networking Release 2. Sun was a powerhouse by that time, not a seedling.
Yawn
Supercomputers also do not provide a complex environment for users - they are for power not for sophistication. Having a simple OS increases the raw power.
Please go and change the Wiki page then as they also seem to be under the mistaken impression that Sun was built on BSD.
Where do I get my information? From working in IT for 40 years and actually using software from its infancy. It's hard to be so impressed with Apple when you saw the Xerox interface they copied or tried to use their Apple Lisa when it came out.
But keep reading Rudy's history, you obviously like fiction a lot ;-)
11 processor architectures
How does that affect your argument?
Appears, none at all
Exactly (NT)
11 processor architectures
It doesn't
Really? How do you know that?
How do you know that they are inefficient? It is very hard to do an apples to apples comparison to verify your claim that Linux doesn't work well on ARM for example.
The sheer proliferation of embedded linux seems to indicate that whatever it might lack in efficiency it is certainly good enough.
I think the Linux being tied to x86 thing is overblown.
EEMBC.org
Not that they go out of their way to tell you, of course -check the compilers! :)